AI Tool Reduces Unexpected Hospital Deaths by 26%, Canadian Study Reveals

A recent study conducted by researchers at Unity Health Toronto and published in the Canadian Medical Association Journal highlights the efficacy of CHARTWatch, an AI-based early warning system developed to monitor patients at St. Michael's Hospital in real-time. This system, which identifies individuals at high risk of unexpected death or ICU transfer, has been shown to reduce unanticipated mortality by a remarkable 26% in the hospital's general internal medicine ward.

"AI tools are becoming more prevalent in healthcare, but they must be evaluated for safety and effectiveness," said Dr. Amol Verma, lead author of the study and a clinician-scientist at Unity Health. Dr. Verma, who spearheaded the development and implementation of CHARTWatch, emphasized that their findings underscore the potential of AI in reducing unexpected hospital deaths.

The study reviewed data from over 13,000 admissions, examining outcomes on the general internal medicine ward against those from subspecialty units where CHARTWatch was not deployed. The team found no reduction in unexpected deaths in units not using the AI tool, affirming its direct impact on patient survival.

AI Tool Reduces Hospital Deaths by 26%, Canadian Study Finds | by Shubham | Sep, 2024 | Medium

"Seeing this result solely in the units with CHARTWatch is a promising sign," Verma noted, adding that the AI-driven system was meticulously developed based on input from medical staff. The technology analyzes over 100 data points, such as vital signs and lab results, to provide dynamic hourly risk assessments.

Clinical nurse educator Shirley Bell shared how CHARTWatch alerted her team to a patient's severe infection, cellulitis, hours before routine checks would have revealed the issue. This early warning allowed the team to administer life-saving antibiotics, underscoring how CHARTWatch enhances rather than replaces traditional care.

According to Dr. Muhammad Mamdani, vice president of data science at Unity Health, CHARTWatch exemplifies how AI can augment medical care by supplementing clinicians' assessments with real-time data insights. This technology operates unobtrusively in the background, continuously assessing patients' conditions and alerting healthcare providers when necessary.

What is Artificial Intelligence (AI) and Why People Should Learn About it - UCF Business Incubation Program - University of Central Florida

CHARTWatch's success is fueling expansion plans. Unity Health's collaboration with GEMINI, Canada's most prominent hospital data-sharing network, aims to test the AI tool across more than 30 hospitals in Ontario, laying the groundwork for broader adoption.

"We're hopeful about CHARTWatch's potential impact beyond our hospital network," Mamdani said. "These early interventions could prove transformative in diverse clinical settings."

Despite its promise, CHARTWatch has limitations. Dr. Verma pointed out that the study occurred at an urban hospital during the COVID-19 pandemic and was not a randomized controlled trial. Dr. John-Jose Nunez from the University of British Columbia praised the study's "real-world" implications but emphasized further research in varied settings to understand the technology's scalability and ensure patient privacy fully.

The Positive Impact of AI in Healthcare (linkedin.com)

In a broader context, AI technologies like CHARTWatch are part of a rising trend in healthcare. From cancer detection to monitoring vital signs for early disease warning, AI tools are increasingly recognized for their potential to bridge gaps in traditional healthcare models, particularly in rural and underserved areas.

Dr. Verma remains optimistic about AI's role in healthcare, envisioning it as a crucial team member complementing medical professionals. "Few AI technologies have reached clinical settings," Verma stated. "CHARTWatch stands out as a pioneering tool that actively supports patient care daily."